The U.S. District Court for the North District of California dismissed four of six claims in a pair of cases alleging that the use by OpenAI, Inc. of the plaintiffs’ books infringed the copyrights in those books. Tremblay v. OpenAI, Inc. and Silverman v. OpenAI, Inc., 2024 U.S. Dist. LEXIS 24618 (N.D. Cal. Feb. 12, 2024). The books in question include The Cabin at the End of the World by novelist Paul Tremblay and The Bedwetter by comedian Sarah Silverman.

Open AI is a Large Language Model (“LLM”)

OpenAI develops and distributes the artificial-intelligence (“AI”) software known as large language models (“LLMs”). OpenAI uses large texts, the “training dataset,” to train its LLM ChatGPT, which copies text from the training dataset and extracts “expressive information.” ChatGPT can then respond to users’ queries by simulating human reasoning to answer questions or summarize books. OpenAI copied the plaintiffs’ books and used them in its training dataset, enabling ChatGPT to generate accurate summaries of the books’ content and themes.

The Plaintiffs Represent a Class

The plaintiffs seek to represent a class of all people in the United States who own a copyright in any work used in an OpenAI training dataset. The plaintiffs asserted six causes of action against various OpenAI entities:

(1) direct copyright infringement (Count I);

(2) vicarious infringement (Count II);

(3) violation of Section 1202(b) of the Digital Millennium Copyright Act (“DMCA”) (Count III);

(4) unfair competition under Cal. Bus. & Prof. Code Section 17200 (“UCL”) (Count IV); (5) negligence (Count V); and

(6) unjust enrichment (Count VI).

The defendants moved to dismiss Counts II-VI under Federal Rule of Civil Procedure 12(b)(6) for failure to state a claim upon which relief can be granted.

With respect to Count II, the court noted that “[a] claim of vicarious infringement requires a threshold showing of direct infringement” (citing Perfect 10, Inc. v. Amazon.com, Inc., 508 F.3d 1146, 1169 (9th Cir. 2007) (other citations omitted)). The court found insufficient the plaintiffs’ allegation that “every output of the OpenAI Language Models is an infringing derivative work.” According to the court, the plaintiffs failed to explain what the ChatGPT outputs entail or to “allege that any particular output is substantially similar—or similar at all—to their books.” Thus, the court dismissed the vicarious copyright infringement claim with leave to amend.

Count III rested on Section 1202(b) of the DMCA, which provides that one cannot, without authority, (1) intentionally remove or alter any copyright management information (“CMI”), (2) distribute CMI knowing that it has been removed or altered, or (3) distribute copies of works knowing that the works’ CMI has been removed or altered. OpenAI argued that (i) the plaintiffs’ claim under Section 1202(b)(1) failed because they did not plausibly allege that OpenAI intentionally removed CMI during the training process or intended to conceal or induce infringement and (ii) their claim under Section 1202(b)(3) failed because they did not allege that the defendants distributed the copyrighted works or copies of them.

The court held that the plaintiffs made only conclusory allegations that “[b]y design, the training process does not preserve any CMI.” The court also held that the plaintiffs did not allege that the defendants distributed the plaintiffs’ books or copies of their books. Instead, the court said, they “alleged that ‘every output from the OpenAI Language Models is an infringing derivative work’ without providing any indication as to what such outputs entail—i.e., whether they are the copyrighted books or copies of the books.” The court concluded that the plaintiff’s allegations were “insufficient to support this cause of action under the DMCA” and therefore dismissed the DMCA claims with leave to amend.

Regarding Count IV, alleged unfair competition, the court noted: “A business act or practice may violate the UCL if it is either ‘unlawful,’ ‘unfair,’ or ‘fraudulent.’ Each of these three adjectives captures ‘a separate and distinct theory of liability’” (quoting Rubio v. Capital One Bank, 613 F.3d 1195, 1203 (9th Cir. 2010) (citing Kearns v. Ford Motor Co., 567 F.3d 1120, 1127 (9th Cir. 2009)). The court said that although the plaintiffs alleged that the defendants engaged in “unlawful business practices” by violating the DMCA, the court had dismissed the predicate DMCA claims and thus the derivate UCL claim of unlawfulness could not survive. In addition, the court said that because the plaintiffs failed to satisfy the heightened pleading requirements for UCL fraud claims, that claim also could not survive.

With respect to the question of unfairness, the court said that “[a]ssuming the truth of Plaintiffs’ allegations—that Defendants used Plaintiffs’ copyrighted works to train their language models for commercial profit—the Court concludes that Defendants’ conduct may constitute an unfair practice.” Thus, the court held that portion of the UCL claim could proceed.

The court dealt quickly with the claims of negligence and unjust enrichment. Because the plaintiffs did not allege that the defendants owed them a legal duty, the court dismissed the negligence claim with leave to amend. And because the plaintiffs did not allege “that OpenAI unjustly obtained benefits from [the plaintiffs’] copyrighted works through fraud, mistake, coercion, or request,” the court said, that claim also failed.

Accordingly, the court granted in part and denied in part the defendants’ motions to dismiss. The court ordered that the amended complaint consolidate the claims in Tremblay and Silverman with those in the similar case of Chabon v. OpenAI, Inc., 2024 U.S. Dist. LEXIS 24618 (N.D. Cal. Feb. 12, 2024).

This decision shows that courts may not be sympathetic to “tangential” claims of copyright owners in AI cases (other than possibly unfair trade practices). It remains to be seen how sympathetic courts may be to claims of direct infringement through unauthorized reproduction of copyrighted works or possibly through other violations of the exclusive rights in copyrighted works set forth in Section 106 of the Copyright Act.